zhi chen
2015-Apr-21 05:57 UTC
[LLVMdev] what's meaning of the last field of an LLVM intrinsic instrucation
The intrinsic instructions always have an number in the end, what's meaning of it? For example, %2 = call <2 x double> @llvm.x86.sse2.sqrt.pd(<2 x double> %1) #2 What does this #2 mean? Can I ignore it when I create the instruction? Thanks. Best, Zhi -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150420/c61ea7bb/attachment.html>
Robinson, Paul
2015-Apr-21 06:21 UTC
[LLVMdev] what's meaning of the last field of an LLVM intrinsic instrucation
The #2 refers to a set of attributes. Attribute sets are given unique numbers. You should see them listed in the IR, somewhere after the functions. I believe attributes are attached after the instruction is created. They generally apply to functions (and calls), and not to most instructions. --paulr From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] On Behalf Of zhi chen Sent: Monday, April 20, 2015 10:57 PM To: LLVM Dev Subject: [LLVMdev] what's meaning of the last field of an LLVM intrinsic instrucation The intrinsic instructions always have an number in the end, what's meaning of it? For example, %2 = call <2 x double> @llvm.x86.sse2.sqrt.pd(<2 x double> %1) #2 What does this #2 mean? Can I ignore it when I create the instruction? Thanks. Best, Zhi -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150421/69e23cac/attachment.html>
zhi chen
2015-Apr-21 06:29 UTC
[LLVMdev] what's meaning of the last field of an LLVM intrinsic instrucation
Hi Paul, Is this field mandatory? or do I have to create the attribute if I create a call for the intrinsic function? Thanks, Zhi On Mon, Apr 20, 2015 at 11:21 PM, Robinson, Paul < Paul_Robinson at playstation.sony.com> wrote:> The #2 refers to a set of attributes. Attribute sets are given unique > numbers. You should see them listed in the IR, somewhere after the > functions. > > I believe attributes are attached after the instruction is created. They > generally apply to functions (and calls), and not to most instructions. > > --paulr > > > > *From:* llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] *On > Behalf Of *zhi chen > *Sent:* Monday, April 20, 2015 10:57 PM > *To:* LLVM Dev > *Subject:* [LLVMdev] what's meaning of the last field of an LLVM > intrinsic instrucation > > > > The intrinsic instructions always have an number in the end, what's > meaning of it? > > > > For example, %2 = call <2 x double> @llvm.x86.sse2.sqrt.pd(<2 x double> > %1) #2 > > > > What does this #2 mean? Can I ignore it when I create the instruction? > Thanks. > > > > Best, > > Zhi >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150420/ec5d8b1c/attachment.html>
Seemingly Similar Threads
- [LLVMdev] what's the best way to insert an instruction after the current instruction
- [LLVMdev] what's the best way to insert an instruction after the current instruction
- error of using GATHER intrinsic
- how to force llvm generate gather intrinsic
- how to force llvm generate gather intrinsic